Research at the Adaptive Communications and Signal Processing Group (ACSP) at Cornell University focuses on three major areas:

Adaptive techniques for wireless networks

Wireless communication and information theory

Adaptive signal processing

We investigate theoretical and practical issues in wireless networks including large scale sensor networks and high performance wireless LAN. We are interested signal processing and information theoretic aspects of the physical layer design and the interaction between the physical layer and the medium access control layer. Applications include future wireless systems, wireless LAN, ad hoc networking, terrestrial HDTV broadcasting, and DSL/cable modems.

Current projects

• Signal Processing for Random Access: A Cross Layer Approach (NSF)

This research aims to develop novel random access techniques for two emerging wireless applications: high data-rate wireless LAN for hot-spot access and information extraction in large-scale and low-power sensor networks. For both applications, conventional medium access control that assumes simplistic channel model is one of the major performance bottlenecks. We explore new cross-layer design strategies by exploiting tight interactions between signal processing at the physical layer and the medium access control sublayer. The research has two major themes. The first involves developing optimal medium access strategies that capitalize the multi-packet reception capabilities of the physical layer and provide capacity achieving random access. Signal processing techniques are developed to maximize the stable throughput of users. Rate allocations, quality of service requests, and pricing structures are designed based on the characterization of the capacity and stability region of optimal random access protocols. The second theme focuses on low power and energy efficient random access for large-scale sensor networks. The cross-layer design approaches take the form of utilizing channel state information at the physical layer. Distributed, scalable random access protocols with low complexity are investigated for the sensor reach-back problem. Novel signal processing techniques at the sensor node and the mobile access point are developed for energy efficient medium access.
 
• Channel and QoS Adaptive Multimedia Wireless Ad-Hoc Networks (ONR)

Funded by the Department of Defense, this Multidisciplinary University Research Initiative (MURI) focuses on new design methodologies for wireless multimedia networks. Specific topics include signal processing for wireless networks, protocol aided signal processing for adaptive receivers, adaptive source and channel codings for multimedia wireless networks, and asynchronous VLSI circuit simulation of large scale wireless networks.
 
• Wireless Ad-Hoc Networks with Receiver Multipacket Reception

Supported by the Army Research Office (ARO), this project investigates receiver multipacket reception (MPR) in wireless ad-hoc networks. The objective of the research is twofold. First, a general theoretical framework is developed for the analysis of code division multiple access (CDMA) ad-hoc networks whose nodes are capable of simultaneously detecting multiple packets. Potential improvements due to MPR in network throughput, delay, and stability are measured against the increased complexity of receiver implementations. Tradeoffs among spread spectrum bandwidth, error control, and packet error probability are examined. Second, multipacket reception techniques are developed and analyzed for slow frequency hopping code division multiple access (SFH-CDMA) ad hoc networks. Packet formats, spatial and propagation diversities are exploited for signal processing based receiver MPR techniques. Resolvability of multipacket reception techniques is analyzed as a function of signal-to-noise ratio, multiaccess interference, and transmission protocols.
 
• Adaptive Signal Processing in Asynchronous Wireless Networks

This project is supported by the National Science Foundation. The objective is to develop adaptive and spectrally efficient channel equalization and estimation schemes for wireless ATM and other asynchronous networks. Methodologies that do not rely on the transmission of training signals will be considered. Emphasize will be placed on the exploitation of data packet structure, medium access control protocols and temporal and spatial diversities. This includes the possible use of ATM cell header information, the guard-period in the multiple access protocol, and the error correction and detection structure imbeded in the ATM cels.
 
• Channel Estimation and Equalization in Packet Radio Networks

This project is supported by the Center for Research on Applied Signal Processing (CRASP). The thrust of the research is transceiver designs for packet radio networks. Specifically, we investigate blind and semi-blind channel estimation and equalization in packet transmissions such as those employed in EDGE, GPRS, Bluetooth and Wideband CDMA.
 
• Adaptive Receivers for Terrestrial Broadcast HDTV

This project focuses on adaptive receiver design for the next generation terrestrial broadcast HDTV. The research includes adaptive channel estimation and equalization techniques for broadcast channels. This research is supported by Philips.